Most engineers think…
Most people think the FTD access control policy is just a firewall ACL: a list of allow/deny lines, top to bottom, done. That mental model is half-right and it is exactly why people misuse Trust and accidentally shadow their own rules.
The Access Control Policy (ACP) is the master policy on FTD, and it sits inside a bigger pipeline. Before a rule is even evaluated, the prefilter runs in LINA, then optional decryption, then Security Intelligence. Only then do the rules run top-down, first match, and the action you pick — especially Trust vs Allow — decides whether Snort ever inspects the traffic. Understanding that order is what stops you from trusting traffic you meant to inspect.
① Security zones & rule anatomy — what a rule matches on
A security zone groups one or more interfaces — say all the inside interfaces into an INSIDE zone and the internet-facing one into an OUTSIDE zone. Zones are how a rule talks about where traffic comes from and goes to without naming raw interfaces.
An access control rule then matches on a set of criteria: source/destination zones, networks (objects), ports, applications (Layer-7 app identity, not just ports), URLs (category or specific), and users (from identity). A rule can use one criterion or several — all listed criteria must match for the rule to fire.
The interview line: zones plus L3/L4 plus L7 plus identity — an FTD rule is far richer than a classic 5-tuple ACL, which is exactly why the action you attach to it matters so much.
What is a security zone used for in an access control rule?
② Rule actions — Block, Allow, Trust, Monitor
The action decides the fate of matched traffic. Block silently drops it; Block with reset drops and sends a TCP reset. Allow permits it and hands it to Snort — so an Allow rule can run an intrusion policy and a file/malware policy for deep inspection. Trust permits the traffic but skips deep inspection entirely — no IPS, no file inspection. Monitor never decides the fate of traffic; it only logs the connection and lets evaluation continue to lower rules. Interactive Block warns the user with a page they can click through.
The gotcha you must name: Trust vs Allow
Both let traffic through, but only Allow inspects. If you Trust traffic you meant to inspect, Snort never sees it — malware or exploits ride straight through. Use Trust only for traffic you are deliberately exempting from inspection (e.g. a trusted high-volume backup), and use Allow with an intrusion policy for anything you actually want protected.
A logical group of interfaces used as source/destination match criteria, so rules say INSIDE→OUTSIDE instead of naming raw interfaces.
Allow permits and inspects — it can run intrusion and file policies in Snort. Trust permits but skips deep inspection entirely.
Runs first in LINA, before Snort: fastpath, block or analyze, plus tunnel rules. Cheap early decisions for traffic that needs no deep inspection.
Reputation-based IP, URL and DNS blocking from Cisco Talos, applied before the access control rules run.
Say it cleanly in interviews: Allow permits AND hands traffic to Snort for intrusion and file inspection; Trust permits and skips inspection. If you want IPS or malware checks on a flow, it must be Allow — reserve Trust for traffic you are deliberately exempting.
Which rule action permits traffic but skips intrusion and file inspection?
③ The prefilter and the full inspection order
Not everything reaches Snort. The Prefilter policy runs first, in the LINA data plane, before deep inspection. Its actions are Fastpath (permit and skip the rest — cheapest), Block, and Analyze (send on to access control / Snort), plus tunnel rules for encapsulated traffic. It exists to make cheap, early decisions on traffic that does not need deep inspection.
The full order, end to end
- Ingress (LINA) — interface, stateful checks, connection table.
- Prefilter — fastpath / block / analyze, tunnel rules.
- Decryption — optional TLS/SSL policy so the rest can see inside HTTPS.
- Security Intelligence — reputation IP / URL / DNS block (Talos).
- Access Control rules — top-down, first match.
- Snort — if the match is Allow, run intrusion & file inspection.
- Egress — forward the permitted, inspected packet.
Trust short-circuits this: a Trust match permits and skips the Snort stage entirely.
A common slip is forcing every decision into Snort-level access control rules. Cheap, high-volume traffic that needs no deep inspection belongs in the prefilter (fastpath) or Security Intelligence — making those early LINA-stage decisions keeps Snort free for traffic that actually needs inspection.
▶ Watch a packet run the access control gauntlet
How one packet is evaluated top-down against the ACP. Press Play for the healthy inspected path, then Break it to see the classic failure.
In the FTD inspection order, what happens first, before Snort?
④ Default action, rule order & best practice
Rules are evaluated top-down, first match wins: the first rule whose criteria all match decides the action, and evaluation stops (Monitor is the exception — it logs and keeps going). Anything that matches no rule falls through to the Default Action at the bottom of the policy.
The shadowing trap
Because it is first-match, a broad rule above a specific one shadows it. Put a wide Trust any-any near the top and the specific Allow-with-IPS rule below it never runs — traffic is trusted and Snort never inspects it. The fix is ordering: specific rules above broad ones.
The interview line: order specific before broad, prefer Allow-with-IPS over Trust for anything you want inspected, and let the default action be an explicit block — not an accident.
Arjun at a Chennai SaaS firm faces this
He built an Allow rule with an intrusion policy for traffic to the app servers, but the IPS dashboard shows zero events for that path even under a known test exploit.
A broad Trust any-any rule sits above the specific Allow-with-IPS rule, so the app traffic matches Trust first and bypasses Snort.
First-match means the higher Trust rule wins; the specific Allow rule is shadowed and never evaluated, so no intrusion inspection ever happens.
FMC ▸ Policies ▸ Access Control ▸ rule orderMove the specific Allow-with-IPS rule above the broad Trust rule (or scope the Trust narrowly), then deploy.
Re-run the test exploit — the intrusion dashboard now logs events on that path, confirming Snort is inspecting the Allowed traffic.
Do not guess which rule a flow hits. Use the rule hit counts and connection events in FMC to see exactly which rule matched. If a specific rule shows zero hits while a broader rule above it climbs, you have found a shadowed rule — fix the order and re-check.
A broad Trust any-any rule sits above a specific Allow-with-IPS rule. What happens to that specific traffic?
🤖 Ask the AI Tutor
Tap any question — instant, scoped to this lesson. No login, no waiting.
Pre-curated from vendor docs + community Q&A, scoped to this lesson. For a live prod issue, paste your export into chat.techclick.in.
📝 Wrap-up assessment — six more
You've answered 4 inline. Six left. 70% (7 of 10) marks the lesson complete on your profile. Tap Submit all answers at the end.
🧠 In your own words
Type one line: why can a broad Trust rule above a specific Allow-with-IPS rule let an exploit through? Then compare with the expert version.
🗣 Teach a friend
Best way to lock it in — explain it in one line to a teammate. Tap to generate a paste-ready summary.
📖 Glossary
- Access Control Policy (ACP)
- The master policy on an FTD — an ordered set of rules plus a default action that decides what traffic is allowed, blocked, trusted or monitored.
- Security zone
- A logical group of one or more interfaces used as source/destination match criteria in access control rules.
- Allow
- Permits matched traffic and hands it to Snort, so an intrusion policy and a file/malware policy can run on it.
- Trust
- Permits matched traffic but skips deep inspection entirely — no intrusion or file inspection by Snort.
- Monitor
- Logs the connection but does not decide the fate of traffic; evaluation continues to lower rules.
- Prefilter policy
- An early LINA-stage policy that fastpaths, blocks or analyzes traffic (plus tunnel rules) before Snort deep inspection.
- Security Intelligence
- Reputation-based IP, URL and DNS blocking using Cisco Talos intelligence, applied before the access control rules.
- Default action
- The catch-all applied to traffic that matches no rule — e.g. block all, or an intrusion-inspected allow.
- Shadowed rule
- A specific rule placed below a broader one that matches first, so the specific rule never runs under first-match evaluation.
📚 Sources
- Cisco — Secure Firewall Management Center Device Configuration Guide: Access Control Policy overview and rules. cisco.com
- Cisco — FTD: security zones and interface objects. cisco.com
- Cisco — Access control rule actions — Block, Allow, Trust, Monitor and Interactive Block. cisco.com
- Cisco — Prefilter policies and tunnel rules (fastpath, block, analyze). cisco.com
- Cisco — Secure Firewall: traffic processing / inspection order — LINA, Security Intelligence, access control and Snort. cisco.com
- Cisco Talos — Security Intelligence feeds and Snort 3 intrusion rule sets. talosintelligence.com
What's next?
Got the policy and the order? Next, go deep on Snort 3 intrusion policies — rule groups, base policies, the Cisco Talos rule set, and how an Allow rule hands traffic to deep inspection without slowing the LINA data plane.